Hi, I report a bug I can always reproduce with the ati driver using xrandr 1.2. I own a t43p with a ati firegl 3200. I plug my external monitor and the xrandr -q show me the external display. next I do xrandr --output VGA-0 --auto to clone the display to the external monitor. Then the right of the screen is black for 3-4 centimeters and the bottom of the desktop is hidden. (I can only see half of the taskbar in kubuntu which is 48 pixel). All the rest is ok. I noticed it since 6.6.191 (I don't remember before) and even with 6.7.194. I will uplaod my xorg.conf and xorg log, and will providing pictures by the end of the week
Created attachment 11728 [details] xorg.conf
Created attachment 11729 [details] [review] Xorg.0.log
Can you also attach the output of xrandr?
Created attachment 11748 [details] result of xrandr -q
There it is, I seen nothing weird in it. Is there a more verbose output ?
(In reply to comment #5) > There it is, I seen nothing weird in it. > Is there a more verbose output ? > xrandr --verbose will show you more. BTW, the output you posted only shows one output active (as noted by the *). I just wanted to make that's what you were seeing at the time as well.
Created attachment 11749 [details] xrandr --verbose, both screens connected This is what I get from xrandr --verbose (screen size is correct for vga-0), Any clues on what could cause this ? - Wrong output settings ? (BTW the received inputs seems ok)
Does disabling tiling help? Option "ColorTiling" "FALSE"
(In reply to comment #8) > Does disabling tiling help? > Option "ColorTiling" "FALSE" > I just add Option "ColorTiling" "FALSE" in my xorg.conf and it doesn't change anything ?
Created attachment 11820 [details] Picture of the problem You can see that the desktop is not fully displayed, on the right a few pixels are missing, and the bottom is cut at the middle of the taskbar. Tell me if you wat a more accurate/descriding picture.
Do you still have this problem with the latest code from ati git?
I didn't try, I now run the 6.7.195 from ubuntu packages and the problem is the same. Where is the git repo ? Thanks for all your work.
Hem, I just did that : git-clone git://anongit.freedesktop.org/git/xorg/driver/xf86-video-ati cd xf86-video-ati ./autogen.sh make make install Is that right, to get the bleebing edge code? Howerver the above didn't change anything. How can I be sure that my card is able to have the first monitor at 1400x1050 and the second at 1024 768 or above ?
(In reply to comment #13) > Hem, I just did that : > git-clone git://anongit.freedesktop.org/git/xorg/driver/xf86-video-ati > cd xf86-video-ati > ./autogen.sh > make > make install > Is that right, to get the bleebing edge code? Yes. > Howerver the above didn't change anything. > How can I be sure that my card is able to have the first monitor at 1400x1050 > and the second at 1024 768 or above ? it should work fine. you monitor may just not like that particular mode. try: xrandr --output VGA-0 --mode 1280x1024 --rate 59.9 or xrandr --output VGA-0 --mode 1280x1024 --rate 75.0
Ok, now I did this : xrandr --output VGA-0 --auto It's the same result (black stipe on the right and missing part of the bottom ) and xrandr --output VGA-0 --mode 1280x1024 --rate 59.9 gives something interesting : For the bottom it doesn't change anything, but for the right part, it changes something. Instead of a black stripe, the missing part is out of the screen, the output and been streched to cover all the screen. But the same size of pixels is missing.
can you try with --auto and use the controls on your monitor to adjust the image? Also, what about the 75 Hz mode?
Hehe, @ home where I perform the test, I have a buttonless LCD screen. It's a samsung 193p, but I had the same issue with a crt at my university. In the settings are done with --auto nornally. I will test with the 75hz (LCD support 75Hz ?)
(In reply to comment #17) > Hehe, @ home where I perform the test, I have a buttonless LCD screen. > It's a samsung 193p, but I had the same issue with a crt at my university. > In the settings are done with --auto nornally. Ok. > I will test with the 75hz (LCD support 75Hz ?) > Presumably it's coming from your monitor's edid. there's a scaler and other equipment in the monitor to convert the incoming signal (within reason) to the panel's native mode.
(In reply to comment #10) > Created an attachment (id=11820) [details] > Picture of the problem > > You can see that the desktop is not fully displayed, on the right a few pixels > are missing, and the bottom is cut at the middle of the taskbar. > Tell me if you wat a more accurate/descriding picture. > I think this is a duplicate of bug 12834. The edid sync polarity parsing was broken in xserver 1.3. you need to patch your xserver or manually use a corrected mode.
(In reply to comment #19) > (In reply to comment #10) > > Created an attachment (id=11820) [details] [details] > > Picture of the problem > > > > You can see that the desktop is not fully displayed, on the right a few pixels > > are missing, and the bottom is cut at the middle of the taskbar. > > Tell me if you wat a more accurate/descriding picture. > > > > I think this is a duplicate of bug 12834. The edid sync polarity parsing was > broken in xserver 1.3. you need to patch your xserver or manually use a > corrected mode. > Hi Alex I took time to test against xserver 1.4, since I'am not quite able to patch recompile xserver. However I just install a dual boot with hardy heron on it, which provides xserver 1.4.1 (git 20071212). I just test xrandr --auto and it doesn't change anything the same black stipe appears on the left while something is missing on the bottom of the screen. I actually use 6.7.197 from git, can I help in any way ? Since I only have this laptop with an ATI hardware I can't test the driver with different hardware. Thanks for all the good work you do.
Can you post your new log with xserver 1.4 and ati 6.7.197?
Created attachment 13292 [details] [review] xorg.conf / xserver 1.4 and ati 6.7.197 Here is the xorg.conf generated by the ubuntu installer, I just add the option tiling FALSE, as suggested.
Created attachment 13293 [details] [review] Xorg.0.log / xserver 1.4 ati 6.7.197 The text around the line 1606 to the bottom is produced by connecting the second screen (1280x1024) and running xrandr -q. I wonder what this means ? RADEON(0): Supported Future Video Modes: (II) RADEON(0): #0: hsize: 1280 vsize 1024 refresh: 60 vid: 32897 (II) RADEON(0): #1: hsize: 1152 vsize 864 refresh: 75 vid: 20337 Does it mean that there is a limitation if I use two displays, in terms of size of output.
(In reply to comment #23) > Created an attachment (id=13293) [details] > Xorg.0.log / xserver 1.4 ati 6.7.197 (II) ATI: ATI driver wrapper (version 6.7.196) for chipsets: mach64, rage128, radeon you're still using an old version of the driver. > > The text around the line 1606 to the bottom is produced by connecting the > second screen (1280x1024) and running xrandr -q. > I wonder what this means ? > > RADEON(0): Supported Future Video Modes: > (II) RADEON(0): #0: hsize: 1280 vsize 1024 refresh: 60 vid: 32897 > (II) RADEON(0): #1: hsize: 1152 vsize 864 refresh: 75 vid: 20337 > > Does it mean that there is a limitation if I use two displays, in terms of size > of output. > No those are just modes listed in the edid for the monitor you connected.
Oups I was sure to use the 6.7.197 because man ati gives me 6.7.197. I just did git clone as above and then ./autgen.sh make make install. How can I be sure to use the right version ? And how can I be sure to read the latest info in Xorg.0.log since there is no date/time at each line ?
Hello, I just took some time to do another test, and I did it through radeon 6.7.195 and xserver 1.3 on ubuntu gutsy. I can fix the problem by downscaling the LVDS display to 1024x768 and doing xrandr --output VGA-0 --auto && xrandr --output VGA-0 --mode 1024x768. I did this in two step because of the wrong modes that reports xrandr with xserver 1.3. I did this test because when I have tested fglrx 7.12 and the clone mode (same resolution) option was available but not the mirror mode (different resolution). Is it a limitation of the hardware, since fglrx and radeon seem to be unable to do some mirror mode ? I will try the last git code on hardy later... Cheers
(In reply to comment #26) > Hello, I just took some time to do another test, and I did it through radeon > 6.7.195 and xserver 1.3 on ubuntu gutsy. > > I can fix the problem by downscaling the LVDS display to 1024x768 and doing > xrandr --output VGA-0 --auto && xrandr --output VGA-0 --mode 1024x768. > > I did this in two step because of the wrong modes that reports xrandr with > xserver 1.3. what wrong modes? the edid parsing bug? > I did this test because when I have tested fglrx 7.12 and the clone mode (same > resolution) option was available but not the mirror mode (different > resolution). > Is it a limitation of the hardware, since fglrx and radeon seem to be unable to > do some mirror mode ? Shouldn't be a hw limitation. Each crtc can scan the same chunk of framebuffer with it's own timings. Can you try with ati git? Can you also attach your recent log and xrandr output?
Created attachment 13936 [details] [review] log of xorg, with ati from git, on xserver 1.4.0.90
Created attachment 13940 [details] xrandr -q with external diplay plugged. Here are the xorg log, and the output of xrandr. It's still the same, if I do leave the default 1400x1050 on LVDS, and I try to use the external monitor at the default 1280x1024, the bug appears. Then if I downscale the LVDS at 1024x768 (no 1280x1024 mode for LVDS) the bug goes out if I do xrandr --output VGA-0 --mode 1024x768. Then I tried another thing : with the LVDS at 1024x768, and xrandr --output VGA-0 --auto (1280x1024), the display the has a few more bugs. The mouse on the LVDS can out of the screen, i.e in the area of the external display which is not displayed in the LVDS. And some part of the external display are drawn twice (like the gnome panel, even if it's more some bits of memory that are left here instaed of being cleaned). The gnome taskbar is drawn at 768 on both display instead of 768 for lvds and 1024 for external display. So I have the gnome taskbar that hides a bit of the external display. While applications are drawn until the end of the external screen. (and so out of the LVDS).
(In reply to comment #28) > Created an attachment (id=13936) [details] > log of xorg, with ati from git, on xserver 1.4.0.90 > Can you attach the log with the monitor attached? does the following combination work? xrandr --output LVDS --off xrandr --output VGA-0 --mode 1024x768 xrandr --output VGA-0 --auto xrandr --output LVDS --auto Does switching between XAA and EXA help (Option "AccelMethod" "EXA")? How about in combination with Option "ColorTiling" "false"?
> Can you attach the log with the monitor attached? Do you mean by starting X with the second monitor plugged ? (sorry I am not native english speaking, so some I can lose some subtlety in the process) > does the following combination work? > > xrandr --output LVDS --off > xrandr --output VGA-0 --mode 1024x768 > xrandr --output VGA-0 --auto > xrandr --output LVDS --auto > > Does switching between XAA and EXA help (Option "AccelMethod" "EXA")? How > about in combination with Option "ColorTiling" "false"? > I try it now
Created attachment 13949 [details] xorg log with external plugged in Wow, lots of news to say, with the external diplay plugged in. I didn't made the colortilling options now I will test that after. Here is the test : Start normally : (second monitor out) Plug the monitor and restart X. The LVDS starts at 1400x 1050 while the other one starts at 1280x1024 with missing part on the bottom and missing the black stripe. Then I switch off LVDS : The black stripe is stile there but, this time the desktop is fully displayed. (the bottom isn't missing and the right part that was missing is displayed at the left of the black stripe). So the desktop is a little bit streched in. Then I do xrandr --output VGA-0 --mode 1024x768 and now the desktop is fully displayed and use the whole screen. (perfect but not at the right resolution) Then I do xrandr --output VGA-0 --auto and the black stripe is back, but the bottom is still displayed. Then I do xrandr --output LVDS --auto and the bottom goes out of the external display. Next step: redo all of this but with the suggested options.
> Next step: redo all of this but with the suggested options. > Exactly the same effect with those options.
is this still an issue with a more recent version of the driver? Also note that in clone mode dissimilar modes on each head may result in certain areas being cut off if one of the screens is smaller.
I'm seeing this problem with pretty new version of the radeon driver as well (the latest from Ubuntu Jaunty as of today). (II) Loading /usr/lib/xorg/modules/drivers//radeon_drv.so (II) Module radeon: vendor="X.Org Foundation" compiled for 1.5.99.3, module version = 6.10.0 Module class: X.Org Video Driver ABI class: X.Org Video Driver, version 5.0 In my case I have a dual screen setup where the VGA monitor is on the left and my laptop screen is on the right. When I switch the resolution of the monitor to 1600x1200 and the laptop screen to 1400x1050 then the right 3 inches of my laptop screen becomes black and hides the desktop. I can move my mouse into the black area and see it, but i can't see the X desktop at all. This same setup had no problems using Ubuntu Intrepid. Is there anything I can do to help debug the problem?
Hi Alex, As far as I remember, it was the case early in the year, but now the pc in question doesn't run anymore. So I'm unable to test it. Don't now if this bug should be closed or not. Bye, see you in other bug tickets. I'm experimenting with xorg-edgers ppa. Lucas
Closing as INVALID; only reporter no longer has working hardware, and nobody could successfully reproduce the problem.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.